Jan 2018 – Jan 2022 · 4 yrs

• Creator of the LexID Digital with the core linking algorithm, architect for the LexID Digital Confidence Score and Trust Score. Initiator of the Trust Score Model, providing modeling requirement and overseeing the production deployment and model governance.
• Initiator of LNRS ThreatMetrix Flagship Models, managing the modeling and releases of global and regional industry consortium models through model governance and certification process.
• Supervise research and innovation work throughout the key ML products: Digital ID, Smart Learning, Smart ID, TMX Behavior Biometrics, and streaming analytics. Extend advanced research into GNN, LLM/Transformer application in fraud modeling, and other production-focused analytics and modeling components.
• Set ML team goals and directions and R&D strategies, write research proposals, modeling requirements and technical designs, work with the team on prototyping and milestone reviews, promote team initiatives with Product, Services, and Sales teams, produce candidates with proven value for production implementation.
• Mentor the team in continuous learning and innovation. Cultivate an environment that encourages collaboration across teams, create a supportive system empowering individuals for success. Build a team culture of curiosity, resilience, execution and excellence.

Jan 2016 – Jan 2018 · 2 yrs

• Direct the digital identity intelligence research and machine learning product design and implementations. Create innovative solutions and lead proof-of-concept projects in delivering big data driven threat intelligence into the company key product - Dynamic Decision Platform and Customer Facing Analytics.
• Participate in Machine Learning product strategy discussion and roadmap design. Coordinate and provide technical guidance for Product, Services, Pre-sales and Sales on promoting our analytics offerings.
• Design and maintain the ThreatMetrix Digital ID (LexID Digital) core algorithms and analytics modules. Build algorithms for generating accurate and predictive confidence score and trust score for ThreatMetrix Digital ID.
• Provide technical leadership and guidance for "Smart Learning as a Service" modeling procedures, equip services team and customers with the capability of making policy optimization and help onboard more customers to data driven machine learning models.
• Lead the design and implementation of “Anomaly Detection as a Service” to monitor KPI, key customer data attributes and operational anomalies, generate alerts for both internal and external customers. Lead the OPs initiative “Canary Analyzer” for intelligent software releases and statistical behavior analysis to reduce the risk in production.

Jan 2015 – Jan 2016 · 1 yr

• Build real-time analytic solutions to improve Netflix service availability and reliability.
• Design the new Automated Canary Analysis algorithms implemented in Python to improve the statistical test accuracy on the continuous deployment platform.
• Key contributor to the Shared Anomaly Detection library in Python to provide anomaly detection through REST service.

Jan 2012 – Jan 2015 · 3 yrs

• Technical lead and key contributor to Falcon deliverable models. Supervisor in new initiative research and patent feasibility studies. Inventor of Multi-Layer Self-Calibrating analytic models and originator of Falcon 6 Behavior Sorted List technology. Overseer for core technology governance in model architecture and features.
• Collaborate with cross-functional teams in resolving key client issues. Participate in onsite modeling POC for new clients in emerging market. Validate new modeling components and make recommendations for next generation payment card risk models. Directly involve in project kickoffs and presales events.
• Managing consortium credit modeling team of highly talented scientists. Improve team engagement and grow in innovations and collaborations. Ensure modeling agility through efficient project planning. Always 100% on time deliveries. Falcon Credit revenue increases more than 20% consecutively under my supervision.

Jan 2003 – Jan 2005 · 2 yrs

• Doctoral researcher on NSF Grant “ITR: Algorithms and Software for Knowledge Acquisition from Heterogeneous Distributed Data”. Design and implement incremental algorithms for learning from distributed data sources.
• Doctoral researcher on NSF research initiative “Algorithms and software for learning classifiers from ontologies and partially specified data”. Build ontology aware classifiers that learn from data at multiple level of abstractions.